I'm getting internal server error when requesting transcript

[Evertt]

The request takes a very long time, even when I'm giving it a tiny audio file and using the tiny.en model, and it eventually always ends in an internal server error.

Could it be that the micro-service only accepts audio files in some subset of file formats / codecs? If yes, which ones?

I've tried .mp3 and .wav so far.

I was hoping that running the tiny.en model locally in my docker environment would return a result faster than using OpenAI's whisper API, but that does not seem to be the case.

[matusstas]

Hi @Evertt can you please tell me how long is the audio and show me the internal server error?

[Evertt]

can you please tell me how long is the audio

I've tried 2 audio files, one of 5 seconds or so, and one of 40 seconds. Both in mp3 format and wav format they were no more than just a couple of kilobytes.

show me the internal server error?

I don't understand, how would that help you? It's not like it gives any more information other than "internal server error". Are you saying that I should be able to get a more detailed error with a stack-trace from somewhere? If yes, where should I be able to find that?